AI & Machine Learning
Virtual Laboratories
A practical learning environment for data exploration, machine learning, model evaluation, neural networks, NLP, computer vision and responsible AI. Learn by changing inputs, running experiments, comparing outcomes and explaining evidence.
Experiment library
Choose a lab to launch its interactive workspaceData Exploration & Image Classification
Explore a dataset, inspect features, train a transparent k-NN classifier and evaluate predictions.
AI problem scoping
Define a human-centred problem and success criteria.
The 4Ws problem canvas
Investigate Who, What, Where and Why before designing a solution.
Data acquisition
Compare data sources, sampling and collection quality.
Data exploration
Inspect distributions, patterns, missing values and outliers.
AI ethics and responsibility
Explore privacy, fairness, accountability and human oversight.
AI project cycle
Navigate the stages of an AI project.
Data types
Distinguish numerical, categorical, text, image and other data.
Data visualisation
Choose and interpret charts for different questions.
Natural language processing
Explore how machines process human language.
Computer vision
Understand image representation and visual features.
Neural network playground
Trace inputs, weights, layers and activation.
Supervised learning
Learn from labelled examples and compare task types.
Unsupervised learning
Discover groups and structure without labels.
Reinforcement learning
Explore agents, actions, rewards and policies.
Training and testing data
Separate training and testing data to estimate generalisation.
Classification
Predict categories and inspect decision boundaries.
Regression
Predict numerical values and inspect fitted relationships.
Confusion matrix
Interpret true and false predictions by class.
Accuracy and error
Calculate accuracy and understand its limitations.
Bias and fairness
Investigate how data and design choices can create bias.
Python dataset studio
Explore structured datasets and basic data workflows.
Data preprocessing
Clean, encode, scale and prepare model inputs.
Features and target variables
Separate predictors from the value a model learns to predict.
K-nearest neighbours
Classify new samples from nearby labelled examples.
Decision trees
Follow feature-based splits to reach a prediction.
Linear regression
Fit a line and examine residual error.
K-means clustering
Assign points to clusters and update centroids.
Overfitting and underfitting
Compare model complexity, training error and test error.
Precision, recall and F1
Calculate and compare classification metrics.
End-to-end ML pipeline
Connect problem framing, data, modelling and evaluation.
Image classifier concepts
Explore image classification and confidence outputs.
Sentiment analysis
Classify text sentiment and examine language features.
Ethics and deployment
Consider testing, privacy, safety and deployment risks.
Neural network fundamentals
Explore weighted sums, activation and learning concepts.
Model monitoring
Monitor performance, drift and changing data.